Cannot pause pool when streaming data with mysql-node - mysql

I use node and the mysql package to stream data from node to client.
The idea is,
define a pool, and queries based on the pool.
Then pass the streaming rows to an array.
If that array's length reaches a length, pause the stream, process the rows, send them to client via websockets.
Resume stream. Repeat until no other rows are left.
I am following the examples on the mysql npm page but I get pool.pause is not a function
Here is the code
var pool = mysql.createPool({
connectionLimit : 100,
host : config.host,
user : config.user,
password : config.password,
database : config.database
});
//turn simple queries to promises
const query = (str, ar) => {
return new Promise(function(resolve, reject) {
pool.query(str, ar, function (error, results, fields) {
if (error) {
return reject(error);
}
resolve({results, fields});
});
})//promise
}
const userdetails = (ws, data) => {
//do a check, unrelated to streaming
query('SELECT COUNT(id) as countrows FROM users WHERE category = ? ', [data.category])
.then((data)=>{
if(data.results[0].countrows > 5000){
// if more than 5000, we stream
// the following is based on the mysql code found in their page
// it has no relation to the promise-based query above
var query = pool.query('SELECT id, name, address, sale, preexisting, amount FROM users WHERE category = ? ', [data.category])
query.on('result', row => {
rowsToProcess.push(row);
if (rowsToProcess.length >= 100) {
pool.pause();
processRows();
}
});
query.on('end', () => {
processRows();
});
const processRows = (done) => {
//process some data
//send them back using websockets
ws.send(JSON.stringify({ data }));
pool.resume();
}
}
})
}
I dont know if this is related to making a simple query , a promise or using the pool, or anything else. This gives the TypeError: pool.pause is not a function and I cannot fix it. Please advice.
Thanks

You can try this solution,
I have used this many times:
const mysqlStreamQueryPromise = (queryString, params) => {
return new Promise((resolve, reject) => {
let streamData = connection.query(queryString,params).stream();
let data = [];
streamData.on('data', item => {
streamData.pause();
data.push(item);
streamData.resume();
});
streamData.on('end', end => {
return resolve(data);
});
streamData.on('error', error => {
return reject(error);
});
});
}

Use this
var pool = mysql.createPool({
connectionLimit : 100,
host : config.host,
user : config.user,
password : config.password,
database : config.database
});
//turn simple queries to promises
const query = (str, ar) => {
return new Promise(function(resolve, reject) {
pool.query(str, ar, function (error, results, fields) {
if (error) {
return reject(error);
}
resolve({results, fields});
});
})//promise
}
const userdetails = (ws, data) => {
//do a check, unrelated to streaming
query('SELECT COUNT(id) as countrows FROM users WHERE category = ? ', [data.category])
.then((data)=>{
if(data.results[0].countrows > 5000){
// if more than 5000, we stream
// the following is based on the mysql code found in their page
// it has no relation to the promise-based query above
var query = pool.query('SELECT id, name, address, sale, preexisting, amount FROM users WHERE category = ? ', [data.category]).stream();
query.on('result', row => {
rowsToProcess.push(row);
if (rowsToProcess.length >= 100) {
pool.pause();
processRows();
}
});
query.on('end', () => {
processRows();
});
const processRows = (done) => {
//process some data
//send them back using websockets
ws.send(JSON.stringify({ data }));
pool.resume();
}
}
})
}

Related

convert a NodeJS lambda function (AWS) to use "async" (promises, etc.) instead of callbacks

I have a lambda function that connects to mysql and runs a set of queries, but I actually have a sequence of mysql queries that need to run one after another. I.e., the value of one query is used in the next query, etc.
Currently, I have a bunch of callbacks to achieve this, but this is leading to "callback hell". How would I rewrite this to use async / await?
My code is actually split into 2 files. The first file does an initial query, and then the value is passed into a function of the second file. Please note that the mysql node_module is included but not shown here. The AWS API gateway calls index.js
// index.js
var mysql = require('mysql'); // from node_modules
var config = require('./config.json');
var dashboard = require('./dashboard.js');
var pool = mysql.createPool({
host : config.dbhost,
user : config.dbuser,
password : config.dbpassword,
database : config.dbname
});
exports.handler = (event, context, callback) => {
context.callbackWaitsForEmptyEventLoop = false;
pool.getConnection(function(err, connection) {
// check for mysql connection error first
if ( err ) {
throw err;
}
let qry = "select id from some_table where some_field = ?";
let someval = event.queryStringParameters.someval;
connection.query(qry, [someval], function(error, result) {
if ( error ) {
throw err;
}
else {
dashboard.processRequest(connection, callback, event, res[0].id);
}
});
});
}
// dashboard.js
module.exports = {
jsonResponse: function(results) {
return {
"statusCode": 200,
"body": JSON.stringify({ results }),
"isBase64Encoded": false,
"headers": {
"Access-Control-Allow-Origin": "*"
}
};
},
processRequest: function(connection, callback, event, val) {
let qry = "update first_table set some_field = ?";
connection.query(qry, [val], function(error, results) {
// return to client if error
if (error) {
callback(null, this.jsonResponse(error));
}
else {
// assume that this table must be update AFTER the previous statement
qry = "select id from second_table where some_field = ?";
connection.query(qry, [val], function(error1, results1) {
// return to client if error
if ( error1 ) {
callback(null, this.jsonResponse(error1));
}
qry = "update third_table set some_field = ? where id = ?";
connection.query(qry, [results1[0].id], function(error2, results2) {
// release connection when all queries are completed
connection.release();
if ( error2 ) {
callback(null, this.jsonResponse(error2));
}
else {
callback(null, this.jsonResponse(results2));
}
});
});
}
});
}
};
It was suggested to me that something like the below code might work. Unfortunately, it does not. I was curious to know why using try...catch blocks in the way shown below is not working, and is it the same thing as what you've shown, but just written differently?
// index.js
var mysql = require('mysql'); // from node_modules
var config = require('./config.json');
var dashboard = require('./dashboard.js');
var pool = mysql.createPool({
host : config.dbhost,
user : config.dbuser,
password : config.dbpassword,
database : config.dbname
});
exports.handler = (event, context, callback) => {
context.callbackWaitsForEmptyEventLoop = false;
pool.getConnection(function(err, connection) {
// check for mysql connection error first
if ( err ) {
throw err;
}
let qry = "select id from users where username = ? limit 1;";
let username = event.queryStringParameters.username;
try {
let res = await connection.query(qry, [event.queryStringParameters.username]);
dashboard.processRequest(connection, callback, event, res[0].id);
} catch (err) {
console.log(err);
}
});
}
// dashboard.js
module.exports = {
jsonResponse: function (results) {
return {
"statusCode": 200,
"body": JSON.stringify({results}),
"isBase64Encoded": false,
"headers": {
"Access-Control-Allow-Origin": "*"
}
};
},
processRequest: async function (connection, callback, event, val) {
let qry = "update first_table set some_field = ?";
try {
let results = await connection.query(qry, [val]);
qry = "select id from second_table where some_field = ?";
try {
let results1 = await connection.query(qry, [val]);
qry = "update third_table set some_field = ? where id = ?";
try {
let results2 = await connection.query(qry, [results1[0].id]);
connection.release();
callback(null, this.jsonResponse(results2));
} catch (error2) {
callback(null, this.jsonResponse(error2));
}
} catch (error1) {
callback(null, this.jsonResponse(error1));
}
} catch (error) {
callback(null, this.jsonResponse(error));
}
}
};
We need use promises.
Typically I follow this approach:
Create one async method mainProcess and have bunch of methods step by step called with in that method. one after the other with await or all at once.
Each async method getConnection and runQuery in this case, called within mainProcess must a Promise.
If any errors from these methods i.e promise rejects from individual methods, goes in catch block of mainProcess().
If no errors, all methods within mainProcess gets executed and goes to then block of mainProcess()
Your code can be refactored like this (just wrote in an editor untested)
var pool = mysql.createPool({
host: config.dbhost,
user: config.dbuser,
password: config.dbpassword,
database: config.dbname,
});
exports.handler = (event, context, callback) => {
context.callbackWaitsForEmptyEventLoop = false;
/**
* Main Lambda Process
*/
const mainProcess = async () => {
// Get Connection
let connection = await getConnection();
// Run Step 1
let qry1 = "select id from some_table1 where some_field = ?";
const response1 = await runQuery(connection, qry1, { someFiledValue: 1222})
// Run Step 2
let qry2 = "select id from some_table2 where some_field = ?";
const resonse2 = await runQuery(connection, qry2, { someFiledValue: 1222})
return 'All Good';
});
}
mainProcess()
.then(result => {
// All lambda success messages are returned from here
callback(null, result);
})
.catch(error => {
// All lambda errors thrown from here
callback(error);
});
};
function getConnection(qry, parms) {
return new Promise((resolve, reject) => {
pool.getConnection(function (error, connection) {
if (error) {
// return to client if error
reject(error);
} else {
// Return response
resolve(connection);
}
});
});
}
/**
* Runs a query, either resolves or rejects
*/
function runQuery(connection, qry, parms) {
return new Promise((resolve, reject) => {
connection.query(qry, [val], function (error, results) {
if (error) {
// return to client if error
reject(error);
} else {
// Return response
resolve(result);
}
});
});
}
When you're dealing with a lambda function which performs an async task you have two solutions:
you can use non async handlers, in which case you need to invoke "callback" on promises as you did in your example
you can use async handlers, which does not requires the "callback" input and that allows you to write async/await code, like the following example:
const mysql = require('mysql2/promise');
exports.handler = async(event, context) => {
//get path variable
const { pathVar } = event.pathParameters;
// get connection
const connection = await mysql.createConnection({
host : process.env.RDS_HOSTNAME,
user : process.env.RDS_USERNAME,
password : process.env.RDS_PASSWORD,
database : process.env.RDS_DB_NAME
});
// get text query
const textQuery = `SELECT field FROM entity WHERE attribute = ${pathVar}`;
// get res
const results = await connection.execute(textQuery);
return {
"statusCode": 200,
"body": results,
"isBase64Encoded": false
}
}
You can have a look at the AWS docs: https://docs.aws.amazon.com/lambda/latest/dg/nodejs-handler.html

Node callback returns mysql result but i cant print to the user with actions on google for dialogflow

I created an intent to get user information based on the ID he provides as param. Using a mysql module i can process the query and get the result. With a callback i can get the result to the main function but the agent ignores once i pass to a conv.ask(). What am i doing wrong?
This is my first script with node. I tried declaring pesquisar_aluno() in a variable so i could use in the main function but it retuns null.
const express = require('express');
const bodyParser = require('body-parser')
const mysql = require('mysql')
const {
dialogflow,
SignIn,
SimpleResponse
} = require('actions-on-google')
app.intent('pesquisar.alunos', (conv, params) => {
const aluno = params.aluno
conv.ask('Vamos pesquisar')
pesquisar_aluno(aluno,function(result){
var resposta = result
console.log(resposta) // returns the result
conv.ask(resposta) // ignores it
})
console.log(resposta) // returns undefined
})
function pesquisar_aluno(aluno,callback)
{
var connection = mysql.createConnection({
host : process.env.MYSQL_HOST,
user : process.env.MYSQL_USER,
password : process.env.MYSQL_PASS,
database : process.env.MYSQL_DB
})
connection.connect()
var query = `SELECT * FROM aluno WHERE id_aluno = "${aluno}"`
connection.query(query, function (error, results, fields)
{
if(error) throw error
var usuario = `RA =>${results[0].id_aluno} Nome => ${results[0].nome}`
if(callback) return callback(usuario)
})
}
Expect conv.ask(resposta) to print the result to the user but its not printing anything
Edit: Changed to promises. It worked!Thanks to Nick Felker and Prisoner
app.intent('pesquisar.alunos', (conv, params) => {
const aluno = params.aluno
conv.ask('Vamos pesquisar')
let nome = pesquisar_aluno_promise(aluno).then(function(results) {
return results[0].nome
}).catch((err) => setImmediate(() => { throw err; }))
return nome.then(function(result){
conv.ask(result)
})
})
async function pesquisar_aluno_promise(aluno)
{
return new Promise(function (resolve,reject) {
var connection = mysql.createConnection({
host : process.env.MYSQL_HOST,
user : process.env.MYSQL_USER,
password : process.env.MYSQL_PASS,
database : process.env.MYSQL_DB
})
connection.connect()
var query = `SELECT * FROM aluno WHERE id_aluno = "${aluno}"`
connection.query(query, function (error, results, fields)
{
if (error) {
return reject(error)
}
resolve(results)
})
})
}
As Nick suggested in the comments, you need to use Promises when you are doing asynchronous operations.
Additionally, however, you need to return that Promise from your Intent Handler so the Intent Dispatcher knows to wait for the result before continuing.
In your case, this can just be adding return, so it might look something like this:
return nome.then(function(result){
console.log(result) //works
conv.ask(result) //should work now
})

Is there a way to pass a value from a mysql callback function to the outer function in express?

I'm using express and npm MySQL to develop an API.I have a json request in this format:
{
"payments":[
{
"PolicyNo": "ME3",
"PaymentDate": "2019-04-16T18:00:00.000Z",
},
{
"PolicyNo": "PIN001q",
"PaymentDate": "2019-04-16T18:00:00.000Z",
}]
}
I want to check the database if the policyNo exists before inserting. To avoid the common ERR_HTTP_HEADERS_SENT, I've looped through the payments querying the database with the PolicyNo. If it exists it's pushed into a success array if it doesn't it's pushed into a failed array.
This works perfectly but I can't access these arrays outside the callback.
Here's what I've tried:
router.post('/bla', (req, res)=>{
const values = []
const failedvalues = []
let sql = 'SELECT PolicyNo from pinclientinfo WHERE PolicyNo=?'
req.body.payments.forEach(element => {
connection.query(sql,element.PolicyNo,(err, rows) =>{
if(!err){
if(rows && rows.length > 0){
values.push(element.PolicyNo, element.PaymentDate)
}else{
failedvalues.push(element.PolicyNo)
}
}
})
})
res.json({
failed:failedvalues,
success:values
})
})
Here's the response I'm getting:
{
"failed": [],
"success": []
}
This has some major problems, mostly conceptually.
Firstly the forEach is synchronous will be called payments.length number of times, but the sql query is Asynchronous so it will complete in the future.
I think you are confused between synchronous and asynchronous functions and how they work.
But you can solve this (in your case) atleast two ways.
1) Use the IN syntax and get the array. Iterate over it and do stuff. "SELECT PolicyNo from pinclientinfo WHERE PolicyNo in (...)"
let sql = 'SELECT PolicyNo from pinclientinfo WHERE PolicyNo IN (' + Array(req.body.payments).fill('?').join(',') + ')'
const policies = req.body.payments.map(p => p.PolicyNo);
const values = [];
const failedvalues = [];
connection.query(sql, ...policies, (err, rows) => {
//now check each row..
rows.forEach(element => {
//Not optimized only depicts logic
///do stuff
/// like fill values and failedvalues
if(policies.indexOf(element.PolicyNo) > -1){
values.push(...)
}else{
failedvalues.push(...)
}
});
res.json({
failed: failedvalues,
success: values
})
})
Which will be 1 DB call.
2) The other approach is (not very good) doing multiple db calls and check for count.
let sql = 'SELECT PolicyNo from pinclientinfo WHERE PolicyNo=?'
let count = 0;
req.body.payments.forEach(element => {
connection.query(sql, element.PolicyNo, (err, rows) => {
if (!err) {
if (rows && rows.length > 0) {
values.push(element.PolicyNo, element.PaymentDate)
} else {
failedvalues.push(element.PolicyNo)
}
}
// check If all Complete
count+=1;
if(count === req.body.payments){
//all complete
res.json({
failed: failedvalues,
success: values
})
}
})
})
BUT SERIOUSLY, USE PROMISE. USE ASYNC/AWAIT USE THOSE SWEET LITTLE FEATURES ES6 GIVES YOU
Check out: this post
because connection.query is asynchronous, so return:
{
"failed": [],
"success": []
}
use promise and await you can synchronized resolve mysql data
use Promise.all() you can synchronized resolve list of promise
router.post("/bla", async (req, res) => {
let values = [];
let failedvalues;
let promises = [];
let sql = "SELECT PolicyNo from pinclientinfo WHERE PolicyNo=?";
req.body.payments.forEach(element => {
promises.push(
new Promise(function(resolve, reject) {
connection.query(sql, element.PolicyNo, (err, rows) => {
if (!err) {
if (rows && rows.length > 0) {
values.push(element.PolicyNo, element.PaymentDate);
} else {
failedvalues.push(element.PolicyNo);
}
}
resolve();
});
})
);
});
await Promise.all(promises);
res.json({
failed: failedvalues,
success: values
});
});

Lambda function MySQL result not working on NodeJs 8.10

I have a code in Node 6.10 and it is working...
But If I convert it to Node 8.10 it's not working
var mysql = require("mysql");
var connection = mysql.createConnection({
host: " localhost",
user: "root",
password: "",
database: "parser_db"
});
exports.handler = async event => {
connection.connect();
let response = {
statusCode: 400,
body: { Method: "Invalid", event }
};
var readTable = "SELECT * FROM documents where id = " + mysql.escape(1);
connection.query(readTable, function(err, results, fields) {
if (err) throw err;
else {
response = {
statusCode: 200,
body: { results }
//body: { results }
};
console.log(response);
return response;
}
});
};
Can some one please help me to detect the problem. It is also not working if I do the MySQL query in separate file and return the result set.
Note : If I print the result using console.log(response) instead returning it's
showing the correct result.
The problem is that you are returning response from within the connection.query() callback function. That makes response the return value for the callback function, not the return value for the outer Lambda function.
One way to restructure this code is as follows:
exports.handler = async (event) => {
connection.connect();
return new Promise((resolve, reject) => {
const readTable = `SELECT * FROM documents where id = ${mysql.escape(1)}`;
connection.query(readTable, (err, results, fields) => {
if (err) {
reject(err);
} else {
resolve({statusCode: 200, body: {results}});
}
});
});
};
In addition to #jarmod's answer, You can also use the util.promisify method to promisify connection.query so that you can use the await keyword, to make the code simpler
const util = require('util');
exports.handler = async (event) => {
connection.connect();
const readTable = `SELECT * FROM documents where id = ${mysql.escape(1)}`;
const connQueryPromisified = util.promisify(connection.query).bind(connection);
const result = await connQueryPromisified(readTable);
return {statusCode: 200, body: {results}};
};

How to avoid deadlock in nodejs mysql with a lot of queries?

I have a lot of urls, for every url I call the function load(url), this function parse the html, extracts the needed data and builds a bulk insert query as you can see in my test.js code. The problem is, if I have to many urls (like 100+), I get a Error: ER_LOCK_DEADLOCK from mysql. I tried to use async.queue but this is somehow not working (I don't know why, maybe I am using is wrongly). How can I run many urls + queries one after another, avoiding parallel execution which I think resulted in a deadlock? Even using async.queue results to a DEADLOCK (not always).
test.js
const request = require('request');
const async = require('async');
const pool = require('./database');
const urls = [
'https://www.quora.com/What-is-the-best-way-to-have-delayed-job-queue-with-node-js',
'https://de.wikipedia.org/wiki/Reinhardt-Zimmermann-L%C3%B6sung',
'https://towardsdatascience.com/the-5-clustering-algorithms-data-scientists-need-to-know-a36d136ef68'
]
let load = function(url) {
request({url: url}, function(error, response, html) {
if(!error) {
console.log(html);
/**
* 1. Parse HTML
* 2. Create Array of Values
* 3. Call pool.query(sql, [values], function(error) { ... })
*/
let data = [{}];
let sql = "INSERT IGNORE INTO tbl_test (title, content) VALUES ?";
let values = [];
data.forEach((item) => { values.push(item) });
pool.query(sql, [values], function(error) {
if(error) throw error;
})
} else {
console.log("handle error...");
}
})
}
let jobs = []
/*urls.forEach((url) => {
//jobs.push(load(url)); // --> Works but fails if the urls list is to big -> mysql deadlock error!
jobs.push(function(callback) { callback(load(url)) });
})*/
let q = async.queue(function(task, callback) {
console.log("Task:", task.uri);
callback();
})
q.drain = function() {
console.log('all task completed');
pool.end();
}
urls.forEach((url) => {
q.push({uri: url}, function(err) {
console.log('finished processing ...')
});
});
databse.js
require('dotenv').config();
const mysql = require('mysql');
let pool = mysql.createPool(
{
connectionLimit: 10,
host: process.env.DB_HOST,
port: process.env.DB_PORT,
user: process.env.DB_USER,
password: process.env.DB_PASSWORD,
database: process.env.DB_NAME
}
);
pool.getConnection((err, connection) => {
if(err) {
if(err.code === 'PROTOCOL_CONNECTION_LOST') {
console.log('Database connection lost.')
}
if(err.code === 'ER_CON_COUNT_ERROR') {
console.log('Database has too many connections.')
}
if(err.code === 'ECONNREFUSED') {
console.log('Database connection refused.')
}
if(err.code === 'POOL_CLOSED') {
console.log('Pool is closed.')
}
}
if(connection) {
connection.release()
}
return;
});
module.exports = pool;
I have changed the code to use async.series instead of async.queue, beacuse the tasks would run in parallel in queue (see: https://caolan.github.io/async/docs.html#queue).
test.js
...
let tasks = [];
context.forEach((ctx) => {
tasks.push(function(callback) { load(ctx, callback) });
});
async.series(tasks, function(err) {
if(err) return next(err);
});